Diagnosis: a £2.8M homeware brand with one tired leg to stand on

This is an anonymised, illustrative composite — not a single real client. Brand, figures, and quotes are representative of the founder-led businesses we work with, assembled from the patterns we see most often. It is published to show the shape and depth of a Newroot diagnosis. No real client data appears here.

The situation

A premium homeware brand, founder-led, around £2.8M in revenue with eleven staff. It scaled on a single channel — Meta — run well, by a founder who knew the product cold. That worked until it didn’t. Over seven months the same ad account that built the business got quietly more expensive while spend held flat at £17–19k a month, and nothing downstream had been built to catch the slack. Growth had been flat for two quarters. They hired a £6,000/month agency to fix it; the agency added spend and reporting and did not change the shape of the problem.

What we found

Three numbers told most of the story. Blended return on ad spend had fallen from 4.1 to 1.8over seven months on roughly flat spend — the same customer bought for far more money, sold the same basket. The site converted cold paid traffic at 1.3%, about half of where a brand with this review profile and price point should sit. And a 38,000-person email listwas driving just 9% of revenue against a category norm of 25–30%, with a repeat-purchase rate of 19% where the norm is closer to 30–35%.

The texture underneath the numbers mattered as much as the numbers. Ninety-four percent of paid spend in the last 90 days ran on four ad creatives, the newest five months old — a Meta auction starved of fresh signal. The mobile product page, carrying 78% of traffic, pushed reviews and price below three full scrolls and gated checkout behind an account. The Klaviyo flows were the platform defaults: a welcome, an abandoned cart, nothing post-purchase, nothing for the second order, no winback. The product itself was genuinely good — 4.8 across 2,100+ reviews — but the positioning was interchangeable premium-homeware language you could swap with three competitors and not notice.

The root cause

The diagnosis was not “your ads stopped working.” The business had a one-legged growth model, and the leg was tiring.

The original growth was real but fragile: a sharp founder, a cheap Meta auction, and good creative timing. None of that is a system — it’s a moment that held. When the auction got expensive, as it does for every brand eventually, there was no second engine, no owned demand, and no conversion buffer to absorb the shock. The flat quarters were not a new problem; they were the original model reaching its natural ceiling. That is precisely why “fix the ads” could not work — there was never a second thing to fall back on while the ads recovered.

The agency made it legible. Hired to fix return on ad spend, they did the literal thing and worked inside the ad account — adding spend, broadening audiences. But the leak was not in the ad account. It was in the 1.3% conversion rate and the 19% repeat rate, neither of which an ad manager touches. They were measured on one metric, so they defended that metric while the business declined underneath it. That is the structural failure of single-discipline help on a cross-discipline problem: the person you hired could only ever fix the part they were hired for, and that was not the part that was broken.

The 90-day plan

Prioritised by return on effort and by sequence, so the early phases fund and de-risk the later ones.

Phase 1 (weeks 1–4) — bank the owned revenue, stop the conversion leak.The fastest money was already in the building. Rebuild the Klaviyo flow architecture — welcome, browse-abandon, cart-abandon, post-purchase, second-order, winback, VIP, segmented by purchase history — aiming to lift email’s revenue share from 9% toward 18% in the quarter. Fix the highest-traffic mobile templates: reviews and price above the fold, reassurance in cart and checkout, the forced-account gate removed, targeting paid conversion from 1.3% to 1.8%+. Launch a structured SMS layer on the consenting segment. No new spend, fastest payback, and it cuts Meta dependence before we touch the ads.

Phase 2 (weeks 4–9) — refeed the channel that already works.Stand up a creative engine producing 15–20 tested ad concepts a month — the volume the auction needs — built on the customer-language goldmine in the reviews. Restructure the ad account into a clean prospecting/retargeting split. Sharpen the positioning into one defensible reason to choose, and thread it through ads, product page, and email so the whole funnel says the same thing. Target: blended return on ad spend recovering toward 2.6–3.0 on flat spend.

Phase 3 (weeks 9–13) — start the second leg. Begin one deliberate second channel — on this profile, organic short-form plus a light creator programme, validated with a small test first. Install a single contribution-margin dashboard so decisions get made on profit, not platform return on ad spend. The goal is a started, measured channel — not yet a pillar, but no longer a single point of failure.

Modelled conservatively, Phases 1 and 2 put £34–41k of additional contribution on the table over the quarter from owned channels and conversion alone, before any paid-efficiency gain.

The lesson

This brand did not have a single broken thing to fix. It had a one-legged growth model that had reached its ceiling, and three downstream leaks making the ceiling feel lower than it was. That is not a job for an ad manager, and it was never a job a £6k single-discipline retainer could do — it is cross-discipline by nature.

The order is the whole point. Fix the demand you already pay for before you spend another pound buying more.Bank the owned revenue, stop the conversion leak, then refeed the channel that built the business — and it has room to work again. The diagnosis took fourteen days. Acting on it is the part that compounds.